🤖 Onboarding Topic #2: CS AI Agent

Eugene Pak

Welcome to week #2 in the Customer Service Management app onboarding series! Today, we’re excited to introduce the Customer Service Management app’s AI agent, which teams up with your support agents to deliver faster, smarter customer support.

Learn how you can build AI-first customer support experiences with the AI agent by watching the Loom or reading below!

Today, customers expect instant help which many companies are deploying AI for. However, many AI customer experiences are made up of generic, scripted responses and don’t give customers the help they need, leading to significant customer frustration.

The Customer Service Management app’s AI agent not only provides instant, accurate support by leveraging your organization’s knowledge sources, but you also can configure it to ensure it operates at the same high standards as your support agents.

How it works

• Enable the AI Agent:

  • Navigate to the “Customer Experiences” section in the Customer Service Management app, select the experience you want to turn the agent on for, and turn on the toggle. By embedding in channels like your support site, your customers can ask questions about the problems they have and the AI agent can provide them a response, fueled by your knowledge sources.

• Personalize the Agent:

  • Click on “Identity” on the left-hand side and then “Edit Studio” in the right-hand corner. With the ability to configure the AI agent’s name, description, tone, and conversation starters, you can provide a tailored support experience that always matches your brand from the first message.

• Connect Knowledge Sources:

  • In the “Knowledge” section on the left-hand side, configure the agent to pull answers from Confluence, public websites, or other knowledge bases to ensure accurate, up-to-date responses for your customers.

• Add Guidance and Actions:

  • In the “Actions” section on the left-hand side, define specific responses for sensitive topics and set up actions (like service status checks) so the agent can help beyond just answering questions. To make sure the agent operates as you want it to, you can also preview before publishing.

• Test and Iterate:

  • In the “Conversation review” section, you can monitor resolved, unresolved, and escalated conversations and give feedback on how responses could be improved. That way, you have direct control over AI behavior, not just content, for more tailored customer interactions.

• Automate Handoffs:

  • Set up static or automated handoff messages and forms for seamless escalation to your support engineers.

If you have any questions or comments, please add them down below and we’ll be sure to answer. And stay tuned for the next video which will dive deeper on how you can store and access critical customer context so your support agents always know who they’re helping and can deliver personalized customer support experiences!

See other articles in this series:

• 🏃 Onboarding Topic #1: Configure customer experiences

  • Learn how to configure support sites, forms, articles, branding, and channels

• 📁 Onboarding Topic #3: Build customer context

  • Resolve issues with complete context at your fingertips. Customer data, relevant services, team information, and documentation all live in one workspace

• 🔄 Onboarding Topic #4: Establish support workflows

  • Get your teams up and running to actually process and work on customer support requests

Loh Wai Kit

Hi Eugene,

I would like to ask if the AI agent can be embedded into the email channel. Currently, we are using JSM for our external support and most of our customers submit requests via email. So, I was wondering if the AI agent is able to respond to or handle questions coming from email.